I won’t extrapolate toy models that predict end of human civilisation due to tail risks. None of that is going to happen.
even if the tail risks do happen like ice sheet melting, it is over 1000s of years before it truly makes an impact. And even then the impact is minor in grand scale of things.
You could always find some looney model that predicts the world would end if we don’t control emission. Agree to disagree on that.
The only "looney model" is thinking "AI" poses any existential risk while ignoring the coupling of food production to heavily threatened biophysical systems.